CROSS-REFERENCE TO RELATED APPLICATION
FIELD OF INVENTION
[0002] This invention relates to microelectronics and semiconductor circuitry. More specifically,
the invention relates to charge pump voltage multipliers. Even more specifically,
the invention relates to the reduction of negative effects of fast load changes or
off-chip electrostatic discharge in such multipliers.
BACKGROUND OF THE INVENTION
[0003] Cross-coupled MOS inverter cells, driven by capacitively-coupled complementary clock
signals are efficient building blocks in charge-pumps. These cells may be used to
elevate an input DC voltage to a higher voltage output level. The cells may also be
used to reduce an input DC voltage to a lower voltage output level. A positive input
DC voltage may optionally be reduced to an output level below zero volts.
[0004] Known applications of these cells are proposed in
P. Favrat, P. Deval, M.J. Declercq, "A High-Efficiency CMOS Voltage Double," IEEE
Journal of Solid-State Circuits, Vol. 33, No. 3. March 1998. and
R. Pelliconi et al., "Power Efficient Charge Pump in Deep Submicron Standard CMOS
Technology," Proc. 27 ESSC1RC, 2001. As illustrated in FIG. 1, which is an alternative illustration of Pelliconi's Figure
1 or portions of Figure 2 of
J. Cha, "Analysis and Design Techniques of CMOS Charge-Pump-Based Radio-Frequency
Antenna-Switch Controllers, IEEE Trans. On Circuits and Systems - 1: Regular Papers,
Vol. 56, No. 5, May 2009, these disclosures describe a dual-bucket cell that may act as a voltage doubler.
[0005] As illustrated in FIG. 1, herein, an input voltage Vlow is input to two MOSFET inverters.
The first inverter comprises NMOS transistor M1 and PMOS transistor M3, while the
second inverter comprises NMOS transistor M2 and PMOS transistor M4. Both inverters'
outputs are coupled to output voltage Vhigh. A clock signal clk is coupled via capacitor
C1 to the gates of M1 and M3, and the drains of M2 and M4. Circuitry for generating
a clock signal is not illustrated herein, but many circuits for generating clock signals
are well-known to those of ordinary skill in the art. The inverse of clock signal
clk is represented as inverted clock signal nclk, which is low when clk is high and
vice-versa. Circuitry for generating signal nclk is not illustrated, but is well-known
in the art. The inverted clock signal nclk is coupled via capacitor C2 to the gates
of M2 and M4 and the drains of M1 and M3. One of ordinary skill in the art will recognize
the manner in which the circuitry illustrated in FIG. 1 may output a higher voltage
at node Vhigh than is input at node Vlow.
[0007] FIG. 2 sets forth an example of cascaded dual-bucket cells that may be used for voltage
elevation. As illustrated in FIG. 2, each of cells 205, 207, 210 and any number of
intermediate cells represented by ellipses (...) may be cascaded. Each of cells 205,
207, 210, and any intermediate cells may be configured in the manner of the circuitry
illustrated in FIG. 1. Input voltage V_LOW_IN is input into node 212, which corresponds
to Vlow. Cell 205 receives the input at node 212 and outputs a higher voltage at node
206, which corresponds to Vhigh. Node 206 is coupled to the input Vlow of cell 207.
Cell 207 receives the input at node 206 and outputs a higher voltage at node 208,
which corresponds to Vhigh. Node 208 may be coupled to node 209 or, alternatively,
to the input of an intermediate cell. Node 209 is coupled to the output voltage of
the preceding cell and corresponds to Vlow for cell 210. Cell 210 receives the input
at node 209 and outputs a higher voltage at node 211, which corresponds to Vhigh.
Alternatively, as set forth above, any or all of the cells in the cascade may be configured
to output a voltage that is lower than the input voltage. Thus, the labels V_LOW_IN
and V_HIGH_OUT are representative of a a typical use, but V_LOW_IN may actually be
a higher voltage than V_HIGH_OUT.
[0008] A clock input signal CLK_IN is preferably provided to amplifiers 201, 202. Amplifier
201 outputs amplified clock signal clk to each of cells 205, 207, 210, and any intermediate
nodes (...) via capacitors 203. Amplifier 202 outputs inverted clock signal nclk to
each of cells 205, 207, 210, and any intermediate nodes (...) via capacitors 204.
Capacitors 203 and 204 are not illustrated herein for intermediate nodes (...), but,
if used, will be connected in the same fashion as those illustrated with respect to
cells 205, 207, 210. That is, capacitor 203 for any intermediate nodes (...) will
be connected between signal clk and the node. And capacitor 204 for any intermediate
nodes (...) will be connected between signal nclk and the node.
[0009] Node 211 provides output voltage V_HIGH_OUT from cell 210. Node 211 is preferably
coupled to ground via capacitor 213.
[0010] Cascaded cells of the type illustrated in FIG. 2 may be used to provide a much higher
multiple of the input voltage than single cells of the type illustrated in FIG. 1.
The cascaded circuit of FIG. 2 eliminates at least one of the drawbacks of a Dickson
charge pump, in that it does not result in voltage drops across the diodes that are
present in Dickson charge pumps.
[0011] In the preferred normal operation of the device of FIG. 2, the voltages at the various
nodes 212, 206, 208, 209, 211 along the ladder are relatively regularly distributed
between the voltage at V_LOW_IN and the voltage at VIHIGH_OUT. In addition, some small
capacitors and relatively large resistance across each of the cells (for example,
205, 207, 210) may reduce ripple and irregular effects of leakage. Because of this,
the local voltage difference across each cell 205, 207, 210 will not exceed the relatively
low power supply of the clock drivers. If the well in which the devices are constructed
is designed to withstand relatively high DC voltage, the local MOS transistors within
each cell may be constructed with a gate made of a thin oxide layer. This thin oxide
gate construction generally allows for more compact component size and higher efficiency
for any given on-resistance of a component. As taught by
M.D. Ker, S.L. Chen, C.S. Tsai, "Design of Charge Pump Circuit With Consideration
of Gate-Oxide Reliability in Low-Voltage CMOS processes," IEEE Journal of Solid-State
Circuits, Vol. 41, No. 5, May 2006, this type of thin oxide gate structure is not expected to suffer gate-oxide reliability
problems.
[0012] However, if node V_HIGH_OUT is connected to off-chip circuitry with an external final
decoupling capacitor, it is desirable to make the structure insensitive to electrostatic
discharge ("ESD"). Because the node V_HIGH_OUT is preferably a high voltage node,
it is desirable to use ESD protection devices within the pad. Neither these devices
nor the pad is illustrated herein, because such structures will be familiar to one
of ordinary skill in the art. Yet, even with such structures, the on-chip voltage
at node V_HIGH_OUT may be subjected to sharp and relatively high voltage transients.
These transients may be much higher than the transients encountered in normal chip
operating conditions in a chip having thin oxide devices. In such transients, the
final cell 210 of the charge pump is exposed to the voltage transients and may be
damaged or destroyed by the transient.
[0013] It would be desirable to reduce the amount of damage or destruction to the cells
of the charge pump due to sharp and/or high voltage transients from off-chip sources.
[0014] It would also be desirable to reduce the amount of damage or destruction to other
types of cells sensitive to overvoltage across cells. Accordingly, the invention described
herein is not limited to applicability to cross-coupled MOS inverter cells.
SUMMARY OF THE INVENTION
[0015] The present invention reduces or eliminates the identified problems by providing
novel circuitry and modes of operation for use with cascaded voltage elevation cells.
The present invention reduces or eliminates these problems by providing an embodiment
for coupling a capacitor to nodes vlow and vhigh in parallel with one or more cells
within the cascaded structure.
[0016] In another embodiment, capacitors may be coupled in parallel across more than one
cell within the cascaded structure.
[0017] In yet another embodiment, a resistor may be coupled in series between the final
cell and the output load.
BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The above and other objects and advantages of the present invention will be apparent
upon consideration of the following detailed description, taken in conjunction with
the accompanying drawings, in which like reference characters refer to like parts
throughout, and in which:
FIG. 1 illustrates prior art circuitry for a dual-bucket cell voltage elevator.
FIG. 2 illustrates prior art circuitry for cascaded dual-bucket cells for voltage
elevation.
FIG. 3 illustrates exemplary circuitry according to the present invention.
FIG. 4 illustrates the potential improvement in voltage curves that may be seen through
use of the present invention.
FIG. 5 illustrates exemplary circuitry according to the present invention.
FIG. 6 illustrates exemplary circuitry according to the present invention.
DETAILED DESCRIPTION OF THE INVENTION
[0019] FIG. 3 illustrates an exemplary embodiment of the present invention, including a
capacitor 312 with capacitance C_esd coupled between nodes 209 and 311 to a voltage
multiplier cell 210. A resistor 313 with resistance R_esd may be electrically coupled
between node 311 and node V_HIGH_OUT 314. Finally, a capacitor 315 may be electrically
coupled between node 314 and ground to reduce the effects of electrostatic discharge.
When such a capacitor is absent or is insufficiently large, the effect of electrostatic
discharge may be much more problematic.
[0020] If the small capacitors (mentioned above) that are coupled across cell were large,
these capacitors would allow for better distribution of transient overvoltages across
several cells, thereby reducing the potential for damage to or destruction of individual
cell, including the final cell 210. When ESD is encountered, however, the clock capacitors
203, 204 act as if they were capacitors to ground. The clock capacitors 203, 204 also
attenuate the overvoltage on the innermost cells (e.g., those cells between node 209
and node 206) but leave the final cell 210 exposed to the ESD. Capacitors 203, 204
may, for example, be implemented as 6 pF thick oxide capacitors in a charge pump delivering
30 V with a 50 µA to 100 µA load current. For these reasons, the clock capacitors
are generally incapable of solving the problem that the present invention seeks to
resolve. Additionally, increasing the clock capacitors 203, 204 to a capacitance sufficient
to solve the ESD problem would significantly increase the size of the capacitors 203,
204 and undesirably increase the area needed to form these capacitors on the chip.
[0021] Thus, it is desirable to form capacitors (e.g., capacitor 312) that are sufficiently
large to reduce the impact of the overvoltage resulting from ESD and thereby protect
the final few or final several cells. Two examples of possible configurations of such
capacitors are set forth in FIGS. 5 and 6. A preferred embodiment in FIG. 5 includes
capacitor 312 electrically coupled in parallel with cell 210 at nodes 311 and 522.
Capacitor 519 is preferably electrically coupled in parallel with cell 517 at nodes
521 and 522. And capacitor 518 is preferably electrically coupled in parallel with
cell 516 at nodes 520 and 521. Because the final cell 210 will typically encounter
the largest voltage from ESD, it is preferable to form capacitor 312 with a larger
capacitance than the other capacitors. For example, in an implementation of a charge
pump voltage multiplier that is configured to deliver 30 volts with a 50 µA to 100µA
load current, capacitor 312 may be a 10pF to 15pF capacitor while capacitors 518 and
519 may each be formed as 5pF or smaller capacitors. One of ordinary skill in the
art will recognize that charge pump voltage multipliers of much different (i.e., greater
or lesser) capacity may be implemented using the invention disclosed herein. For example,
in a device targeting a 5 µA load current, the capacitors may have much smaller capacitance
and in a device with greater capacity, having larger MOS devices, the capacitance
would be increased as needed to provide the benefits of the invention. It is recognized
that at some point in scaling the device to a larger capacity, the size of the device
upon a chip would become substantially large and may reach some limits of feasibility
based upon the amount of chip area needed for implementation of the capacitors and
other semiconductor devices. While only three cells 516, 517, 210 are illustrated
in FIGS. 5 and 6, one of ordinary skill in the art will recognize that the configurations
illustrated in these FIGS. may be extended to additional cells.
[0022] As illustrated in FIG. 5, in a preferred embodiment, cell 210 is the final cell and
node 311 is electrically coupled to V_HIGH_OUT, preferably either via a resistor 313
or directly. The vhigh node of cell 210 is coupled to node 311 and the vlow node of
cell 210 is coupled to node 522. Cell 517 directly precedes cell 210. The vhigh node
of cell 517 is coupled to node 522 and the vlow node of cell 210 is coupled to node
521. Cell 516 directly precedes cell 517. The vhigh node of cell 516 is coupled to
node 521 and the vlow node of cell 516 is coupled to node 520. With respect to the
embodiment set forth in FIG. 3, node 520 may correspond to node 212, node 206, node
208, or any intermediate node (represented by "...") that is removed from node 311
by at least three voltage multiplier cells.
[0023] Another embodiment, as illustrated in FIG. 6, may replace the functionality of capacitors
518 and 519 with differently configured capacitors 618 and 619. Capacitor 619 is configured
to span the final two cells 517 and 210 and is coupled in parallel to these cells
at nodes 521 and 311. Capacitor 618 is configured to span the final three cells 516,
517, and 210 and is coupled in parallel to these cells at nodes 520 and 311.
[0024] The configurations set forth in FIGS. 5 and 6 are interchangeable in various fashions.
For example, capacitors 312 and 519 may be combined with capacitor 618 rather than
518. In another exemplary configuration, capacitors 312 and 518 may be combined with
capacitor 619 rather than 519. Alternatively, either of capacitors 518 or 618 may
be eliminated. It would also be possible to use, for example, capacitors 618 and 518
together or capacitors 619 and 519 together to achieve the effects of this invention.
These various configurations will distribute the overvoltage in different manners
that may be desirable based upon differing configurations of the various cells in
the cascade.
[0025] Distributing the overvoltage resulting from ESD over the final two cells 517, 210
or final three cells 516, 517, 210 is often sufficient to reduce or eliminate the
damage or destruction that would otherwise result if the final cell 210 bore the full
overvoltage. The present invention also contemplates that it is possible to distribute
the overvoltage over a larger number of cells by spanning these additional cells with
capacitors. However, such distribution may not be necessary and may result in unnecessary
waste of chip area in the formation of additional capacitors. One of ordinary skill
in the art will recognize that capacitors 312, 518, 519, 618, 619 may be implemented
in many fashions, including metal-metal, MOS, etc.
[0026] In addition to the addition of capacitors (e.g., capacitors 312, 519, 518, 618, 619),
it is also desirable to add a resistor 313 in series between node 311 and V_HIGH_OUT
node 314. This resistor 313 will participate in the attenuation of ESD pulses along
with the capacitors. If used, the additional resistance of resistor 313 will slightly
degrade the normal performance of the voltage elevator. But resistor 313 can be configured
such that this performance degradation is within an acceptable range, because the
output impedance of the described voltage elevator is relatively high. One of ordinary
skill will recognize that various configurations of resistor 313 may be optimal depending
on the goals of the designer, the configuration of the other components, and the designer's
preferences with respect to the acceptable performance degradation versus the desirable
level of protection against ESD pulses.
[0027] FIG. 4 is an exemplary illustration of potential beneficial effects of the present
invention. One of ordinary skill in the art will recognize that with different fabrication
processes or device configurations, even the use of circuit elements with similar
electrical properties will likely result in at least some variance, possibly a large
variance, from the values illustrated herein. Thus, it may not be possible to exactly
replicate the values set forth herein in an experiment. Curve 401 illustrates an exemplary
expected ESD pulse at node V_HIGH_OUT. As illustrated, the pulse of curve 401 peaks
at approximately 30 volts. Curve 402 illustrates the absolute voltage at node 311
without the capacitor 312 of the present invention. As illustrated, curve 402 peaks
at over 24 volts shortly after the peak of curve 401 was reached. Curve 403 illustrates
the absolute voltage at node 522 without the capacitor 519 or 619 of the present invention.
As illustrated, curve 403 peaks at approximately 13 volts. Curve 406 illustrates the
absolute voltage at node 521 without the capacitor 518 or 618 of the present invention.
As illustrated, curve 406 peaks between 2 and 4 volts. The corresponding cells 210
and 517 are subjected to the difference in voltage between the respective input and
output nodes of the cells as illustrated by the curves. Curve 404 illustrates the
absolute voltage at node 311 with the addition of the capacitor 312 of the present
invention. As illustrated, curve 404 peaks at less than 14 volts. Curve 405 illustrates
the absolute voltage at node 522 with the addition of the capacitor 519 of the present
invention. As illustrated, curve 405 peaks at less than 8 volts. And curve 407 illustrates
the absolute voltage at node 521 with the addition of the capacitor 518 of the present
invention. As illustrated, curve 407 peaks at approximately 2 volts. The exemplary
benefits illustrated in FIG. 4 assume an output resistance of 1 kOhm with respect
to curves 402, 403, and 406. An output resistance of 8 kOhm and protective capacitance
of 15 pF is assumed with respect to curves 404, 405, and 407. (As indicated above
with respect to the capacitors used in implementing the disclosed invention, one of
ordinary skill will recognize that a larger resistance may be needed for a lighter
charge pump and a small resistance for a heavy load.) As can be seen, the peak voltage
at node 311 drops from over 24V in curve 402 to less than 14V in curve 404 with the
additional resistor and capacitor in the presence of the 30V ESD pulse illustrated
in curve 401. Similarly, the peak voltage at node 522 drops from over 12V in curve
403 to less than 8V in curve 405 under the same conditions. Such significant reductions
in overvoltage on the final cells will greatly reduce the damage or destruction that
may otherwise be caused by ESD pulses.
[0028] One of ordinary skill will recognize that the ESD protection circuitry illustrated
in FIGS. 3, 5, and 6 and further discussed above may be implemented with respect to
one, two, three, or any number of cells within a cascaded circuit such as that illustrated
in FIG. 2. One of ordinary skill in the art will also recognize that the protection
circuitry of the present invention may also be used to protect the cells against fast
variations of the output load or output voltage that occur in circumstances other
than through ESD.
1. A voltage multiplier cell comprising,
an input node electrically coupled to a voltage multiplier circuit within the cell
and configured to receive an input voltage;
an output node electrically coupled to the circuit and configured to output an output
voltage;
a capacitor electrically coupled to the input node and the output node in parallel
with the circuit and configured so as to distribute to the input node a portion of
an electrostatic pulse received at the output node.
2. The cell of claim 1, further comprising,
a resistor electrically coupled to the output node in series between the output node
and an output of the cell configured for connection to a load.
3. The cell of claim 1, wherein the capacitor is formed using metal-metal or MOS fabrication.
4. A voltage multiplier comprising a plurality of voltage multiplier cells of claim 1,
wherein the voltage multiplier circuit in each cell comprises at least two cross-coupled
MOS inverters, and wherein the plurality of voltage multiplier cells are electrically
coupled in series, such that, with the exception of a last cell of the plurality of
voltage multiplier cells, the output node of each of the plurality of voltage multiplier
cells is electrically coupled to the input node of the immediately succeeding voltage
multiplier cell in the series.
5. The voltage multiplier of claim 4, wherein the plurality of voltage multiplier cells
comprises at least three voltage multiplier cells.
6. The voltage multiplier of claim 4, further comprising,
a resistor electrical ly coupled to the output node of the last cell in series between
the output node and an output of the multiplier configured for connection to a load.
7. The voltage multiplier of claim 4. wherein the capacitor is formed using metal-metal
or MOS fabrication.
8. A method comprising,
distributing a transient overvoltage across a plurality of voltage multiplier cells,
wherein a plurality of capacitors performs the distribution.
9. The method of claim 8, further comprising,
distributing the overvoltage in a manner that prevents damage to each of the plurality
of cells.
10. The method of claim 9, wherein,
the voltage is distributed by at least one capacitor coupled in parallel with each
of the plurality of cells.
11. The method of claim 10, wherein,
each of the capacitors is coupled to an input of the at least one cell with which
the capacitor is couple in parallel,
each of the capacitors is coupled to an output of at least one cell and wherein the
output of the at least one cell is electrically closer to the output of the plurality
of cells than the input to which the capacitor is coupled.
12. The method of claim 8, wherein the step of distributing comprises,
distributing the transient overvoltage across at least three capacitors.
13. The method of claim 12, wherein,
the at least three capacitors are electrically coupled in series with one another,
and each of the at least three capacitors is coupled in parallel with a single voltage
multiplier cell.
14. The method of claim 12, wherein,
the at least three capacitors are each electrically coupled at one end to a single
point,
each the at least three capacitors is coupled on the other end to a different point,
such that the first capacitor is coupled in parallel with a single voltage multiplier
cell, the second capacitor is coupled in parallel with two voltage multiplier cells,
and the third capacitor is coupled in parallel with three voltage multiplier cells.
15. The method of claim 8, further comprising,
attenuating the overvoltage through a resistor connected to an output of the plurality
of voltage multiplier cells.